python - Riak TS Performance Benchmark on Writing Data -

i new riakts , trying compare druid riakts in iot area. want checkout efficiency of writes riakts before choosing product.hence, started off below dry-run.

data taken: aahrus (data having 4.3 million rows) client used: python node: single independent node.

process went smooth. took 6 mins insert 4.3 million rows in batch of 1 million rows.

is expected performance or can tweak ?

short answer yes, can tweak more performance. first note riak ts, while can run single node, designed clustered. adding nodes , directing writes nodes via load balancer spread workload increase write speed.

some other points consider current set increase write speed bit:

1. the python client supports batching writes (not sure if batching using or not). based on experience 100 records per batch seems optimal performance.
2. riak ts replicates data 3 time ha reasons. since using single node replication isn't needed. when create table set n value equal 1. increase write speed.

simple example of creating table n val = 1:

create table watermeterdata (   customer_id       varchar   not null,   meter_id          varchar   not null,   time_stamp        timestamp not null,   water_pressure    double    not null,   gallons_per_hour  double    not null,   total_gallons     double    not null,   fault_condition   boolean   not null,   fault_message     varchar,    primary key(     (quantum(time_stamp, 80, 'd')),      time_stamp, customer_id, meter_id    ) ) (    n_val=1 )